Crosscut

I've been going on for a bit about how a code database can apply to specific aspects of Caterpillar. I've said it before, but the inspiration for the code database comes from Unison (and to a lesser extent Smalltalk). And Unison gets a lot out of the way they implement this, that I haven't mentioned so far.

In Unison, code is content-addressed. Meaning each piece of code is hashed, and referenced via this hash. This approach has a lot of advantages. The Unison documentation already does a good job of explaining that, but over the next few days, I'd like to go over some of those advantages, and how I think they apply to Caterpillar.

And by the way, while Caterpillar doesn't have a code database yet, it already implements the hashing and content-addressing. This has turned out to be useful for implementing interactive programming (as demonstrated by a previous prototype), and I've recently re-implemented it in the current version.